Study of the animal organ physiology can be achieved by performing an experimental ex vivo perfusion system. The addition of a porcine kidney, as a homeostatic organ to our previously developed ex vivo liver perfusion model can be a principal step to achieve a better physiological environment.
The use of ex vivo perfused models can mimic the physiological conditions of the liver for short periods, but to maintain normal homeostasis for an extended perfusion period is challenging. We have added the kidney to our previous ex vivo perfused liver experiment model to reproduce a more accurate physiological state for prolonged experiments without using live animals. Five intact livers and kidneys were retrieved post-mortem from sacrificed pigs on different days and perfused for a minimum of 6 hr. Hourly arterial blood gases were obtained to analyze pH, lactate, glucose and renal parameters. The primary endpoint was to investigate the effect of adding one kidney to the model on the acid base balance, glucose, and electrolyte levels. The result of this liver-kidney experiment was compared to the results of five previous liver only perfusion models. In summary, with the addition of one kidney to the ex vivo liver circuit, hyperglycemia and metabolic acidosis were improved. In addition this model reproduces the physiological and metabolic responses of the liver sufficiently accurately to obviate the need for the use of live animals. The ex vivo liver-kidney perfusion model can be used as an alternative method in organ specific studies. It provides a disconnection from numerous systemic influences and allows specific and accurate adjustments of arterial and venous pressures and flow.
Studying a large animal organ physiology in an artificial environment is challenging. Furthermore, maintaining normal physiological and biochemical parameters for a prolonged period requires a closely monitored perfusion method and model. We have previously described our ex vivo autologous perfused porcine liver model for the study of hepatic physiology and significant progress has been achieved, since we began our liver reperfusion ex vivo model in 2005, in areas such as harvesting technique, understanding hepatic artery anatomical variations1, inflammatory response2,3, histological changes following liver electrolytic ablation (EA)4, and changes in acid base balance during EA5. Unfortunately hyperglycemia and metabolic acidosis require external supplements of insulin and bicarbonate to maintain an accurate physiological environment. Stable blood sugar levels and acid base balance are crucial when studying islet auto-transplantation, especially in the early implantation phase. We have developed a new model incorporating the kidney to clear excess metabolites and improve the environment for the islet transplantation6,7.
Extracorporeal perfusion experiments lasting longer than 24 hr can also be performed, but these require a larger team to perform the experiments8.
Animals used in this study received human care and study protocols were in accordance with the United Kingdom laws. They were euthanized with final exsanguinations during the blood-harvesting procedure according to United Kingdom regulations. Porcine liver and a left kidney were retrieved from animals in the food chain and that would otherwise be discarded. In this way important ethical implications can be overcome and the economic burden is lower than more common in vivo studies. The short duration of liver and kidney perfusion time in this model, lasting from 6-24 hr, has been a major limitation of our studies; this can be improved by minimizing the exposure of the organs to the ischemic injury (cold and warm ischemia).
1. Preparation of the Procedure Equipment, Solutions, and Experiment Setup
2. Harvesting the Porcine Kidney and Liver and Washing the Organs with Preservation Solution
Animals used in this study received human care and study protocols were in accordance with the United Kingdom laws. They were euthanized with final exsanguinations during the blood-harvesting procedure according to United Kingdom regulations.
3. Priming the Circuit
4. Backbench Preparation of Liver and Kidney for the Perfusion
5. Ex vivo Liver-kidney Perfusion
Liver Function
The liver function tests remained stable for the first 6 hr. However, there were significant changes present only from the 19th hour of the 24 hr experiment (P<0.001). Factor V and X were progressively diminished during the whole experiment and significant changes were evident from the 5th hour of perfusion for factor V and from the 3rd hour for the factor X onward, (P<0.05; Figure 2)20.
Renal Function
The overall levels of urea and creatinine concentrations were significantly lower in the liver kidney model and no significant differences were observed over the hours of perfusion (P<0.001). However, in the liver only model these levels were significantly increased from the first hour onward (Figure 3)9.
Glucose
The glucose concentration normalized when the ex vivo porcine liver model was connected with one kidney when compared to liver circuit alone (P<0.001). There was no major difference between pre and immediately post perfusion changes in blood sugar levels9. In this ex vivo porcine liver-kidney model, the amount of insulin required was much lower compared to the liver ex vivo model alone, where insulin infusion was required continuously to control high glucose concentration (Figure 4).
Acid base Balance
Metabolic acidosis was well controlled with the ex vivo porcine liver-kidney model and sodium bicarbonate was given as required to titrate the acidosis (in boluses). Sodium bicarbonate was not required as a continuous infusion, where it was in the liver ex vivo model. Blood gas analysis showed a significant lower levels of pH in the liver kidney circuits than in the liver circuit alone (P< 0.001). The bicarbonate values of the liver alone model was significantly higher compared to the liver kidney model (P<0.01). A significant increase of the bicarbonate levels was noted in liver alone 3 hr post perfusion onward when compared to the bicarbonate levels at the beginning of the perfusion (p<0.05). We demonstrated a similar result for the base excess result (P<0.05) (Figure 4)9. These results demonstrate that the difference in the acid base balance was obviously corrected by the addition of a homeostatic organ (kidney) in the liver kidney circuit.
Cytokine Response
We used this model to study the differences between IL6 and IL8. Five liver-kidney perfusion experiments compared to seven livers alone model were performed. IL6 and IL8 values were increased considerably after the first and second hour respectively compared with the baseline levels and remained high afterwards (P<0.001). There were no significant differences between liver alone and liver kidney models (Figure 5 and Table 2)10.
Table 1. Experiment bolus and infusion medications.
Table 2. Descriptive data of Interleukins 'IL', Interferon 'INF' and Tumor Necrosis Factor alpha 'TNF' according to the groups. This figure is reprinted with permission from the Journal of Artificial Organs6.
Figure 1. Photograph of the circuit for the ex vivo porcine liver-kidney perfusion model (Left) and schematic liver kidney circuit (Right). This figure is reprinted with permission from the Journal of Artificial Organs10.Click here to view larger image.
Figure 2. Box plot graphs showing the biochemical Liver function outcome after 24 hr liver-kidney experiment, data representing the median values and 95% confidence intervals. This figure is reprinted with permission from the Artificial Organs Journal20.Click here to view larger image.
Figure 3. Error Bars graph demonstrating a comparison of the renal function outcome between liver alone and liver-kidney perfusion experiment for 6 hr. Circles represent the mean and Bars represent two standard deviations. This figure is reprinted with permission from the American Journal of Surgery9.Click here to view larger image.
Figure 4. Error Bars Graph demonstrating Glucose level and metabolic parameters in liver alone and liver-kidney perfusion model for 6 hr. Circles represent the mean and Bars represent two standard deviations. This figure is reprinted with permission from the American Journal of Surgery9.Click here to view larger image.
Figure 5. Error Bars illustrating the IL6 and IL8 outcome in liver alone and liver-kidney perfusion model for 6 hr. Circles represent the mean. This figure is reprinted with permission from the Journal of Artificial Organs10.Click here to view larger image.
In our group the experience and the results achieved with the ex vivo perfusion of the liver have been paralleled over the years by concomitant modifications in the retrieval technique adopted. In the last few years our group has achieved considerable experience using an ex vivo porcine model to study the hepatic physiology5,11, new liver ablation techniques2,5, and liver immunology3. Despite the technical challenges derived from an additional organ, corrections of hyperglycemia, electrolytes imbalances and acid base changes were significant7 and have lead the way to more complex experiments that require strict homeostatic conditions.
The addition of parenteral nutrition to an extracorporeal experiment has been used in a perfusion running for 72 hr12. We didn't provide any parenteral nutrition to the ex vivo perfused organs in our experiment due to the shorter perfusion period.
Comparable to our previous liver ex vivo perfusion models13,14, the outcome of our liver ex vivo perfusion model showed similar changes in the liver physiology, biochemical, immunologic, and pathological findings2,3,5,11, to those more expensive in vivo studies and this confirm our model is reliable and economical acceptable model for the study of porcine liver and related pathology.
The warm ischemia time is an important parameter for the organ survival following transplantation. When longer than 30 min it has been associated with delayed graft function15 and primary nonfunction16. Differently from in vivo retrievals, the ex vivo harvesting process necessarily implies longer warm ischemia times as the animal corresponds, at the time of the harvesting, to a "nonheart-beating" donor and not to a "brain-dead" donor with still a good circulation17. In our initial experiments the wash-out of the warm blood with cold perfusion solutions was achieved once the organs were out of the animal and the average warm ischemia time was approximately 19-20 min. However, to correct the biochemical imbalances produced by the lack of homeostatic organs (lungs, kidney) on the arterial blood gases, electrolytes and glucose levels we decided to add to our model a kidney. The start of a program of multiorgan ex vivo perfusion would necessary require shorter warm ischemia times because the harvesting time would be extended by the retrieval of the second organ. If this bias could not be correct, experiments with transplants particularly sensitive to warm ischemia times, in example pancreatic islet18, could not be possible or would produce biased results.
This technique mimics more closely the retrieval process of human transplantations19. The major organ vessels are approached immediately after the laparotomy when the organs are still in situ. The major difference with the our previous technique is that the organ perfusion with cold solutions, aimed to cool the organ and remove the blood to decrease the metabolic cellular processes, starts earlier when the organs are in the animal and not on the bench table. This difference corresponds to a net gain of approximately 10 min, improving the warm ischemia time of the liver and allowing a good time frame for a second organ to be collected effectively.
Future improvements to reduce the warm ischemia time are still possible. While in the current technique the renal vessels are isolated and the organ harvested after the liver-spleen-stomach-pancreas are removed en-bloc, these vessels could be isolated and cannulated immediately after the perfusion of the liver with the cold preservation solution has started. This would decrease the warm ischemia time of the kidney. Furthermore, similarly to what happen with the current techniques of human transplantation, the organs could be bathed immediately after the laparotomy by large amounts of ice that would decrease the organs' temperature while the vessels are isolated.
In summary, the addition of a single kidney to the previously implemented ex vivo liver perfusion model provides an improved physiological and metabolic environment of the reperfused liver and kidney. In particular, the improved hyperglycemia and acid base status, which make this model more reliable and financially acceptable.
The authors have nothing to disclose.
We would like to thank Sarah Hosgood from the Department of Transplantation, University Hospitals of Leicester, for technical support with the porcine kidney in this model.
Epoprostenol sodium (Flolan) | Glaxo Smith Kline | PL10949/0310 | |
Sodium Bicarbonate 8.4% | Polyfusor, Fresenius Kabi Ltd | PL8828/0043 | |
Cefuroxime sodium (Zynacef) | Flynn Pharma Ltd | PL13621/0018 | |
Calcium Chloride 14.7% | Martindale Pharmaceuticals | PL1883/6174R | |
Heparin | LEO Laboratories Limited | PL0043/0038R | |
Sodium taurocholate | Sigma-Aldrich | T4009-25G | |
Insulin (Actrapid) | Novo Nordisk | EU/1/02/230/003 | |
Soltran (preservation solution) | Baxter Health Care | FKB4708G | preservation solution |
Atraumatic centrifugal pump (Bio-console 560) | Medtronic Inc., Minneapolis, Minnesota- United States; custom pack, cardiac surgery division Europe |